Now Reading
Don’t waste cash on a math coprocessor they mentioned;

Don’t waste cash on a math coprocessor they mentioned;

2023-11-13 01:51:00

You don’t want it they mentioned!

Nicely it’s been no secret, however OS/2 6.123 on my PS/2 mannequin 80, is insanely unstable operating easy MS-DOS based mostly video games (massive EXE’s)

And nearly all the time I’d get this enjoyable error:

SYS0037: The system can not write to the write-protected c: drive.

Adopted by a crash attempting to execute code on the high of the reminiscence MAP (ABIOS?)

Unlawful instruction at 0xffffffff

Then ending this system will simply crash OS/2. Very annoying!

My goto check of v86 mode environments is an previous recreation that I loved as a child, 1988’s BattleTech the Crescent Hawks inception.

Infocom’s Battle Tech

It’s a fantastic recreation, that runs on many 8-bit/16-bit programs of the period, and is surprisingly a really nicely behaved MS-DOS recreation. I imply if Home windows/386 VGA machines can run it in a window utilizing the CGA model, absolutely a brilliant early OS/2 2.0 beta (6.123) can run it, proper? Nevertheless I discovered 6.123 to be extremely unstable, and sadly less than the duty.

I attempted to launch BattleTech again and again and had zero success. I couldn’t determine why it was struggling on my mannequin 80 board, the place it runs simply nice on 86Box. What’s going on?

One factor I had stumbled upon was that if I launched an historical Infocom recreation in a DOS field, after which launched BattleTech it had a a lot larger probability of operating. However this didn’t all the time equate to it working. How is launching an previous COM file from the early 80’s excise the ‘satan’ of some 1988 EXE from operating?

IIT 3C87-25

I wasn’t positive however I had this bizarre suspicion that it was that my system was missing a math coprocessor. After I had the mannequin 60 286 board within the PS/2 case I did spring for an 80287, and one factor I discovered is that OS/2 1.0 & 1.21 ran nice. As a matter of truth I feel it ran higher than after I used to have a 386sx-16 after which later a 486SX-20. Now it’s been nearer to 30 years, so I might have a completely false reminiscence of all this, however I wasn’t positive I used to be onto one thing. So whereas procuring round a subscriber supplied me a math coprocessor as they appear to be insanely costly within the UK. I do not know why the 80287 was so low-cost, and no thought the best way to make any form of adapter, however pJok was capable of rating one for tremendous low-cost in his homeland and ship it to the barren wastelands of Scotland. As I used to be wrapping up the SSD G5 enjoyable, the coprocessor arrived, and it was time to put in it!

Notice the purple 80386! It’s what we would name foreshadowing

The PS/2 8580 motherboard is de facto oddly designed with chip orientation getting in each which different route, and the 80387 socket isn’t keyed by pin, so it’s very important to see the notches on the silkscreen. In any other case I simply used compressed air to blow out the socket, and run the reference disk so as to add the processor.

Put in!

The processor was immediately picked up, though I had the crashing difficulty with the BocaRAM/2 reminiscence card once more, which meant I needed to take away the RAM card, re-configure with the mathematics coprocessor, then add the RAM card, reconfigure, then run the util to patch the CMOS so it’d boot up. I actually dislike this RAM card, however 32bit playing cards price way over this complete endeavor price so I’m just about caught with it.

Now let’s evaluate the Landmark scores between the 286/287 and the 386/ITT387

Landmark System Pace Take a look at with the PS/2 mannequin 60 80286/80287

And now the 386:

Landmark System Pace Take a look at with the PS/2 mannequin 80 80386/ITT 80387

The ITT processor is considerably sooner than the previous 80287, which is fairly superb. The system bus is operating at 16Mhz, though being 32bit vs 16bit yielding a virtually 2x in efficiency, though the ITT co-processor is a lot extra environment friendly.

Booting again into OS/2 6.123, and yeah now it simply works! No fussing round, all the things is simply nice.

I’m form of misplaced too, as none of this could require the maths coprocessor, however the outcomes communicate for themselves. I used to surprise as soon as I received some disk pictures for this historical model of OS/2, why didn’t they ship it? Certain that insane struggle with Microsoft on refusing one thing like Home windows on OS/2, and even WLO like Home windows IN OS/2 from being a part of the product killed any hope of operating apps, however this model of OS/2 is already caught within the lure that it could actually run MS-DOS so nicely, regardless of DPMI not being a factor proper now.

As I’d talked about it does run simply effective in 86Box, so what’s the deal? Nicely that lead me to look again at when it did crash I observed an odd string 038600b1

OS/2 6.123 crash display. TRAP 000e

So what does this imply? Nicely wanting again on the CPU let’s attempt to decode a few of it

16Mhz 80386

First, it’s an A80386-16, which actually isn’t that tough to determine it’s a 16Mhz rated 80386. Subsequent is the revision degree, S40344. Looking out round we will find this table:

S40276 - A1 (however most likely 12 MHz as S40277 is 12 MHz)
S40334 - A2
S40336 - B0
S40337 is B0 stepping
S40343 - B1
S40344 is B1 stepping
S40362 - B1 (20 MHz)

So this locations it at on the tail finish of the introductory line of 386 processors. Checking over at pcjs, we discover that there have been fairly just a few extra revisions to the 386.

See Also

And additional that the B1 Errata is definitely fairly substantial. Possibly this is the reason the 386 had such a poor fame for Unix ports within the day, and why it was shunned by CSRG?

As talked about within the notorious 32bit multiply bug, this processor had been examined and was given the ΣΣ mark of approval. There are quite a few points listed with the presence of a math coprocessor, I’ve to surprise if past points for utilizing the total 32bit datapath, if there have been some electrical points with using the total datapath as nicely? Very like an improperly terminated SCSI bus, did the easy presence of the ITT 387 assist with signaling and enhance system stability? Or am I hitting some bizarre bug in 32bit math that’s simulated because of the lack of a coprocessor, that after one is within the system, the operation is carried out on {hardware}, sidestepping your complete difficulty? I’m neither an EE or any good at reversing code, so I actually don’t know.

The date code 751 does imply that this processor was manufactured within the 51st week of 1987.

how historical this CPU is, I’ve opted to order one which was made in 1990, an SX218 or D1 stepping.

Though it hasn’t arrived but, I’ve to surprise if it might make a very huge distinction in 32bit system stability? I’ve to surprise if there was such an enormous delay in OS/2 2.0 due to the early 386 processors having so many defects that it simply added an undue burden to the event, together with the combating between IBM & Microsoft. Whereas it might be fascinating to see the distinction between any of the Microsoft variations of OS/2 2.0, none have surfaced as of but. Which is a disgrace.

Though it’s good to have this ‘mid’ IBM beta of OS/2, it does undergo from the ever so widespread difficulty of not having the ability to run any delivery 32bit executable, so except you could have supply/object information to hyperlink, you’re just about out of luck. The Microsoft Beta 2 instruments are 16bit, so fortunately they run on just about any model of OS/2, they usually ought to have the ability to run underneath Phar Lap 286 as nicely.

Microsoft OS/2 2.0 tee shirt

One factor that did just lately floor on eBay, is a Microsoft tee shirt from their OS/2 2.0 group. With a minor little bit of sleuthing, the Enterprise is from the 1989 ‘hit’ Star Trek V. Possibly I’m an excessive amount of of a nerd to have acknowledged the GIF.

Again some 20+ years in the past after I lived in Miami, I did have a loaded out PS/2 model 80 again then, and I ran AIX on it, as I believed it was actually cool. However it was additionally extremely unstable. I’ve to surprise now if it was a fault of the processor, or the system? Then once more again then I had 6 registered IP’s and naturally my PS/2 was on the web! Though it was additionally the fitting top to double as a standing mouse pad.

So I assume this doubtlessly leaves us with some painful lesson that you just must get the mathematics coprocessor for older programs in the event you plan on operating something apart from DOS/Home windows with a DOS extender. Whereas I do have a PS/2 model of Xenix, I haven’t been capable of dump them but as my Energy Mac doesn’t like NON FAT disks. One factor is for positive, it made an enormous distinction in OS/2. I don’t suppose 16Mhz/6MB of RAM is anyplace close to sufficient to run OS/2 2.00 at any respectable pace so I’ll stick to the a lot lighter 6.123.

Source Link

What's Your Reaction?
In Love
Not Sure
View Comments (0)

Leave a Reply

Your email address will not be published.

2022 Blinking Robots.
WordPress by Doejo

Scroll To Top